Thanks Masatake, I also prefer to use packages provide by them, this will reduce our work.
As I'm trying to use Bigtop as Ambari's default stack, it'll be better if we use bigtop repo just like mpack, it makes sense to use BIGTOP stack name with BIGTOP repository. But add components to Bigtop requires build it from source, provide rpm/deb build scripts, smoke test cases, and puppet manifest, which I'm doing for ClickHouse right now. And, if ClickHouse end up like ELK, I'll remove it at that time. Best Regards, Zhiguo Wu > On Jul 29, 2022, at 17:21, Masatake Iwasaki <[email protected]> wrote: > > Hi Zhiguo, > > Honestly saying, I'm not positive to add the product to Bigtop. > We do not have enough developer's resource even for maintaining current stack. > Since it is non-ASF product driven by dedicated company, it may end up in the > same way with ELK. > > In addition, ClickHouse seems to provide packaging stuff on their own[1] by > using nFPM[2] > and the packages are published[3]. > If you want additional coverage, contributing to ClickHouse itself should be > the first choice. > > [1] https://github.com/ClickHouse/ClickHouse/tree/master/packages > [2] https://github.com/goreleaser/nfpm > [3] https://clickhouse.com/docs/en/getting-started/install/ > > Regards, > Masatake Iwasaki > > On 2022/07/29 15:25, 吴治国 wrote: >> Hi community, as the ClickHouse are becoming more and more popular OLAP >> engine for bigdata, Although it's not related to Hadoop ecosystem, I still >> want to add it to Bigtop stack. >> But I can only take care of it in some distros(mostly CentOS-7, RockyLinux-8 >> and Ubuntu-20 for now), I don't know if this is enough, if not, can I add it >> to Bigtop? Or is there anyone also interested in this? >> Best Regards, >> Zhiguo Wu
